Would defining , for the sake of this discussion, "normal" exposure as being box speed???
This is a can of worms, to be sure!
"Normal exposure" would be made at box speed if you had previously determined that box speed works for you, generally.
It would be made at something other than box speed if you had previously determined that for some reason box speed does not result in proper exposures,
by whatever criteria you have personally decided upon for defining a proper exposure.
I would say that with proper incident metering, box speed should result in proper-enough exposures for the mid tones (while the SBR as captured with that exposure will determine the low and high tones), even with the normal variation between light meters and shutters caused by different designs and the different amounts "slop" that can develop in these items over time and use. This is what box speeds mean, really: Most people can use this number as an EI to get good exposures under most conditions.
Using an in-camera light meter is a different story! I would not even attempt to judge my film rating based on results from a roll shot under normal conditions (i.e. not test shots) with a reflected light meter that reads through the camera lens (with the possible exception of an in-camera spot meter used for precise tonal placement). There is no reason exposure should change unless the light changes, yet these meters will tell you to change exposure for different subjects, even in the same light. They only lead to the best exposure results if everything in the metered area - metering pattern and weighting considered - averages out to a middle grey tonality, and then the lens is opened up an additional half stop on top of it. IME, it is practically never that things in a composition average out to middle grey; they end up within a certain range, usually, but this range probably extends a stop or two on either side of middle grey. In other words, IME, it is "normal" for exposures made with a directly-read in-camera reflected meter to be anywhere from two stops under to two stops over (including "perfect" from time to time).
I think that 90% of the confusion over box speeds and EIs and "normal" this and "normal" that, and such would be eliminated if people simply threw out the batteries for their cameras' meters, or learned to ignore the meters, and properly used an incident light meter. IMHO, no item has done more to confuse the understanding of ASA, ISO/DIN, EI, and exposure in general than the widespread acceptance and use of the in-camera reflected light meter.
For people who are fine with simply passable results a passable amount of the time, box speed and directly-read in-camera meters usually get the job done just fine, which is what the ASA and ISO/DIN are supposed to be good for. However, these people are probably not the types who will critically judge exposure and spend a lot of time trying to get it such that printing becomes more controllable.
I can surely understand the appeal of this certain degree of technical abandon. Personally, I screwed around with all this EI junk and the Zone Zyztem for years before it just got to be a chore
in my general shooting situations. I now prefer box speed and an incident meter, and I like my results just fine...and previously learning expansion and contraction and abnormal exposure techniques allows me to know when to vary from the norm. It is all about having an eye for SBR and knowing how your various processes and materials will interpret that.
I still use custom EIs and tonal placement for many things, but it is no longer my standard way of working.
I am happy with results using both methods. My subject and "approach" determine which methods I use.
Ideally, I use both in a perfect world...and sometimes the world is perfect, though rarely.